WO2024050637A1 - Method and apparatus for drilling and blasting with fragmentation modeling - Google Patents
Method and apparatus for drilling and blasting with fragmentation modeling Download PDFInfo
- Publication number
- WO2024050637A1 WO2024050637A1 PCT/CA2023/051184 CA2023051184W WO2024050637A1 WO 2024050637 A1 WO2024050637 A1 WO 2024050637A1 CA 2023051184 W CA2023051184 W CA 2023051184W WO 2024050637 A1 WO2024050637 A1 WO 2024050637A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- blast
- features
- fragmentation
- block
- data
- Prior art date
Links
- 238000013467 fragmentation Methods 0.000 title claims abstract description 108
- 238000006062 fragmentation reaction Methods 0.000 title claims abstract description 108
- 238000005553 drilling Methods 0.000 title claims abstract description 51
- 238000005422 blasting Methods 0.000 title claims abstract description 19
- 238000000034 method Methods 0.000 title claims description 25
- 239000002360 explosive Substances 0.000 claims abstract description 111
- 238000013461 design Methods 0.000 claims abstract description 62
- 238000005065 mining Methods 0.000 claims abstract description 42
- 238000009826 distribution Methods 0.000 claims abstract description 41
- 238000010801 machine learning Methods 0.000 claims abstract description 11
- 239000011435 rock Substances 0.000 claims description 58
- 239000012634 fragment Substances 0.000 claims description 31
- 238000005259 measurement Methods 0.000 claims description 15
- 238000007781 pre-processing Methods 0.000 claims description 13
- 238000005457 optimization Methods 0.000 abstract description 5
- 238000013480 data collection Methods 0.000 description 17
- 239000002245 particle Substances 0.000 description 10
- 230000015572 biosynthetic process Effects 0.000 description 9
- 230000008569 process Effects 0.000 description 9
- 238000005474 detonation Methods 0.000 description 8
- 229910052500 inorganic mineral Inorganic materials 0.000 description 8
- 239000011707 mineral Substances 0.000 description 8
- 238000012545 processing Methods 0.000 description 8
- 238000012549 training Methods 0.000 description 7
- RYGMFSIKBFXOCR-UHFFFAOYSA-N Copper Chemical compound [Cu] RYGMFSIKBFXOCR-UHFFFAOYSA-N 0.000 description 4
- 229910052802 copper Inorganic materials 0.000 description 4
- 239000010949 copper Substances 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 239000000463 material Substances 0.000 description 4
- 238000003860 storage Methods 0.000 description 4
- 102100036876 Cyclin-K Human genes 0.000 description 2
- 101000713127 Homo sapiens Cyclin-K Proteins 0.000 description 2
- ZOKXTWBITQBERF-UHFFFAOYSA-N Molybdenum Chemical compound [Mo] ZOKXTWBITQBERF-UHFFFAOYSA-N 0.000 description 2
- 230000004913 activation Effects 0.000 description 2
- 238000009412 basement excavation Methods 0.000 description 2
- 229910052948 bornite Inorganic materials 0.000 description 2
- 239000011575 calcium Substances 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 229910052602 gypsum Inorganic materials 0.000 description 2
- 239000010440 gypsum Substances 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 229910052750 molybdenum Inorganic materials 0.000 description 2
- 239000011733 molybdenum Substances 0.000 description 2
- 238000000926 separation method Methods 0.000 description 2
- VWDWKYIASSYTQR-UHFFFAOYSA-N sodium nitrate Chemical compound [Na+].[O-][N+]([O-])=O VWDWKYIASSYTQR-UHFFFAOYSA-N 0.000 description 2
- 239000007779 soft material Substances 0.000 description 2
- 210000003462 vein Anatomy 0.000 description 2
- 239000010878 waste rock Substances 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- PAWQVTBBRAZDMG-UHFFFAOYSA-N 2-(3-bromo-2-fluorophenyl)acetic acid Chemical compound OC(=O)CC1=CC=CC(Br)=C1F PAWQVTBBRAZDMG-UHFFFAOYSA-N 0.000 description 1
- OYPRJOBELJOOCE-UHFFFAOYSA-N Calcium Chemical compound [Ca] OYPRJOBELJOOCE-UHFFFAOYSA-N 0.000 description 1
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 1
- DGAQECJNVWCQMB-PUAWFVPOSA-M Ilexoside XXIX Chemical compound C[C@@H]1CC[C@@]2(CC[C@@]3(C(=CC[C@H]4[C@]3(CC[C@@H]5[C@@]4(CC[C@@H](C5(C)C)OS(=O)(=O)[O-])C)C)[C@@H]2[C@]1(C)O)C)C(=O)O[C@H]6[C@@H]([C@H]([C@@H]([C@H](O6)CO)O)O)O.[Na+] DGAQECJNVWCQMB-PUAWFVPOSA-M 0.000 description 1
- ZLMJMSJWJFRBEC-UHFFFAOYSA-N Potassium Chemical compound [K] ZLMJMSJWJFRBEC-UHFFFAOYSA-N 0.000 description 1
- 230000004931 aggregating effect Effects 0.000 description 1
- 229910052782 aluminium Inorganic materials 0.000 description 1
- XAGFODPZIPBFFR-UHFFFAOYSA-N aluminium Chemical compound [Al] XAGFODPZIPBFFR-UHFFFAOYSA-N 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 229910052791 calcium Inorganic materials 0.000 description 1
- 229910052799 carbon Inorganic materials 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000003776 cleavage reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000004883 computer application Methods 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 239000000470 constituent Substances 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 238000012938 design process Methods 0.000 description 1
- 230000003292 diminished effect Effects 0.000 description 1
- 238000005315 distribution function Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000008676 import Effects 0.000 description 1
- 229910052751 metal Inorganic materials 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000002085 persistent effect Effects 0.000 description 1
- 239000011591 potassium Substances 0.000 description 1
- 229910052700 potassium Inorganic materials 0.000 description 1
- 239000000843 powder Substances 0.000 description 1
- 230000007017 scission Effects 0.000 description 1
- 239000011734 sodium Substances 0.000 description 1
- 229910052708 sodium Inorganic materials 0.000 description 1
- 235000010344 sodium nitrate Nutrition 0.000 description 1
- 239000004317 sodium nitrate Substances 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Classifications
-
- E—FIXED CONSTRUCTIONS
- E21—EARTH OR ROCK DRILLING; MINING
- E21C—MINING OR QUARRYING
- E21C41/00—Methods of underground or surface mining; Layouts therefor
- E21C41/26—Methods of surface mining; Layouts therefor
- E21C41/30—Methods of surface mining; Layouts therefor for ores, e.g. mining placers
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F42—AMMUNITION; BLASTING
- F42D—BLASTING
- F42D1/00—Blasting methods or apparatus, e.g. loading or tamping
- F42D1/04—Arrangements for ignition
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F42—AMMUNITION; BLASTING
- F42D—BLASTING
- F42D3/00—Particular applications of blasting techniques
- F42D3/04—Particular applications of blasting techniques for rock blasting
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V9/00—Prospecting or detecting by methods not provided for in groups G01V1/00 - G01V8/00
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0499—Feedforward networks
-
- E—FIXED CONSTRUCTIONS
- E21—EARTH OR ROCK DRILLING; MINING
- E21B—EARTH OR ROCK DRILLING; OBTAINING OIL, GAS, WATER, SOLUBLE OR MELTABLE MATERIALS OR A SLURRY OF MINERALS FROM WELLS
- E21B2200/00—Special features related to earth drilling for obtaining oil, gas or water
- E21B2200/22—Fuzzy logic, artificial intelligence, neural networks or the like
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/22—Yield analysis or yield optimisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Definitions
- This disclosure relates to mining, and more particularly to a method of drilling and blasting for excavation.
- rock is broken at a formation, then transported to a mill for processing.
- the mill may, for example, grind rock to extract mineral resources, e.g., ore such as copper ore.
- mill throughput may be a limiting step of the mine operation.
- increasing mill throughput may increase the ultimate output of a mining operation.
- Mining throughput may in turn be influenced by characteristics of the input rock. For example, geological or mechanical properties of input rock, or fragment size of input rock may impact throughput. Generally, harder rock and larger fragments may be more difficult to process and may be associated with lower throughput. Softer rock and smaller fragments may be easier to process and may be associated with higher throughput.
- Rock may be broken at a formation by drilling and blasting. For example, holes may be drilled in a formation, and explosives may be loaded in the holes and detonated. Generally, the quantity and strength of explosive used impacts the size of rock fragments produced from a blast. For example, the amount of explosive may depend on the number and spacing of drilled holes, and the amount of explosive in each hole. Fragment size is generally inversely related to the amount of explosive. That is, blasts involving larger amounts of explosive tend to produce smaller rock fragments. However, increasing explosive quantity also increases the cost of drilling and blasting.
- An example mining system comprises: a fragmentation measurement system installed at a shovel for measuring fragment size following a blast operation at a mine; a data store comprising: for each of a plurality of blocks of a mine, values for a plurality of geological features; values for a plurality of blast features based on hole drilling and explosive data for a plurality of blast events; and for individual ones of the blast events, values for a fragmentation distribution characteristic derived from measurements of fragment size resulting from the individual ones of the blast events; and a blast fragmentation model relating the geological features and blast features to the fragmentation distribution characteristic, the blast fragmentation model produced using a machine learning algorithm and operable to generate a predicted value of the fragmentation distribution characteristic from a blast defined by values for the geological features and the blast features.
- An example mining method comprises: for each of a plurality of blasts at a mine: measuring hole drilling and explosive data for a blast at a mine; computing a value of a fragmentation distribution characteristic from measurements of rock fragment sizes resulting from the blast; associating the blast features with a plurality of geological features for a plurality of blocks at the mine; using a machine learning algorithm, generating a blast fragmentation model relating the fragmentation distribution characteristic to the geological features and blast features based on the hole drilling and explosive data, the blast fragmentation model operable to generate a predicted value of the fragmentation distribution characteristic from a blast defined by values for the geological features and the blast features.
- An example computing system comprises: a processor; a blast fragmentation model produced using a machine learning algorithm, relating: (i) a plurality of blast features based on hole drilling and explosive data for blast events at the mine; and (ii) a plurality of geological features for blocks at the mine; to a fragmentation distribution characteristic; computer-readable instructions which, when executed by the processor, cause the processor to receive the plurality of geological features for a block at the mine, and a corresponding set of blast features, and to generate, using the blast fragmentation model, a predicted value of the fragmentation distribution characteristic from a blast defined by the geological features for the block and the corresponding blast features.
- FIG. 1 is a simplified schematic view of a region of a mine
- FIG. 2 is a schematic view of adjacent layers of a geological block model
- FIG. 3 is a schematic view of a block of the mine of FIG. 1 ;
- FIG. 4 is a side view of a mining shovel
- FIG. 5 is a schematic view of a mining data system
- FIG. 6 is a schematic view of a shovel computing device of the mining data system of FIG. 5;
- FIG. 7 is a schematic view of a data collection system of the mining data system of FIG. 5;
- FIG. 8A is a table showing fragmentation measurements from a machine vision system
- FIG. 8B is a table showing properties in the block model of FIG. 2;
- FIG. 8C is a table showing drilling data in the data collection system of
- FIG. 7
- FIG. 8D is a table showing explosive data in the data collection system of FIG. 7;
- FIG. 9 is a schematic view of an operations platform of the mining data system of FIG. 5;
- FIG. 10 is a table showing a pre-processing process at the operations platform of FIG. 9;
- FIG. 11 A is a table showing block-level drilling data
- FIG. 11 B is a table showing block-level explosive data
- FIG. 11C is a table showing merged block-level data defining block features
- FIG. 11 D is a table showing features of a neighbouring block associated with the merged block-level data of FIG. 11 C;
- FIG. 12 is a schematic view of a user interface of a blast design module of the operations platform of FIG. 9;
- FIG. 13 is a flow chart depicting a drilling and blasting method
- FIG. 14 is a diagram depicting importance of features to output of a blast fragmentation model.
- FIG. 1 is a schematic plan view of a region 100 of a mine.
- the mine may be at a formation containing an ore, such as a copper or other metal ore.
- Region 100 is logically divided into a plurality of blocks 102-1 , 102-2, 102-3 (individually and collectively, blocks 102) for blasting and material removal and extraction.
- Mining of region 100 may be conducted stepwise. That is, segments of rock may be sequentially fragmented.
- blast patterns may be defined, each pattern comprising a series of holes 104 drilled into the rock formation.
- Each blast pattern also defines an amount and type of explosives to be loaded into each hole and detonated (referred to as “drilling and blasting”).
- Each blast may produce muck 106, namely, an aggregate of fragmented rock, for removal and subsequent processing at a mill.
- Blast patterns may or may not be co-extensive with blocks 102. That is, a blast pattern may cover a single block, a portion of a single block, multiple blocks in their entirety, or parts of multiple blocks.
- a detailed three-dimensional model may be produced to represent properties of the formation to be mined.
- the block model defines measured and estimated geological characteristics in blocks 102 of region 100.
- the block model may identify rock types, formation geometry, cleavage planes, mineral content, vein locations and the like.
- Region 100 may be mined in horizontal slices referred to as “benches”. For example, region 100 may be horizontally divided into plurality of benches 108- 1 , 108-2, 108-3 (individually and collectively, benches 108). The vertical thickness of benches and locations and spacing of blocks 102 may be defined based on the block model, e.g., on expected locations of mineral veins, expected rock hardness and the like.
- the muck may be removed using excavation machines referred to as shovels. Specifically, shovels dig muck and transfer the muck to vehicles for transportation. For example, rock with sufficient mineral content may be transported to a plant for processing, and waste rock may be transported for disposal. In an example, transportation is by truck. However, other modes of transport are possible.
- the muck may contain rock fragments of different sizes. The fragment sizes may be measured, and based on those measurements, rock fragmentation may be characterized, for example, in terms of fragmentation distribution, i.e. particle size distribution.
- a muck pile may be described by particle size indices.
- Particle size index P50 denotes a mesh size at which 50 percent of particles pass.
- particle size indices P20, P60, P80 denote mesh sizes at which 20, 60 and 80 percent of particles pass, respectively.
- Fragmentation distribution may impact subsequent processing of muck. For example, muck with smaller fragment size distribution, e.g., smaller P50, may be processed more quickly, such that the throughput of a mill processing the muck may be higher.
- rock with lower hardness and more faults or discontinuities may more easily be fragmented into small pieces, as compared to harder rock with fewer faults or discontinuities.
- rock with larger proportional content of soft minerals may more easily be broken into small pieces as compared to rock with smaller proportional content of soft materials.
- Blasts with high explosive energy may be more likely to produce small rock fragments than blasts with lower explosive energy.
- FIG. 2 schematically depicts two benches 108-1 and 108-2 in a block model.
- Benches 108-1 , 108-2 represent vertically adjacent layers or benches of the mine. That is, benches 108-1 , 108-2 represent rock in layers, and the rock represented by bench 108-1 is positioned immediately atop the rock represented by bench 108-2 and may be referred to as the “bench above” bench 108-2.
- Each bench is divided into a plurality of blocks 102. Individual blocks of bench 108-1 may be positioned directly above corresponding blocks of bench 108-2 and may be referred to as “blocks above” the corresponding blocks of bench 108-2.
- block 102-1 of bench 108-1 is positioned directly above block 102-2 of bench 108-2.
- the block model includes data relating to properties of each block.
- the properties may include, for example, location and spatial data, geometry and rock properties (e.g. mineralogical and geological properties).
- Such data may be relevant to fragmentation of a block resulting from a blast. For example, rock with lower hardness and more faults or discontinuities may more easily be fragmented into small pieces, as compared to harder rock with fewer faults or discontinuities.
- rock with larger proportional content of soft minerals may more easily be broken into small pieces as compared to rock with smaller proportional content of soft materials.
- the results of a blast given specific geometry, rock and explosive properties may be predicted based on results of previous blasts. Such prediction may assist with efficiency of drilling, blasting and removal activities, and subsequent processing of removed material.
- FIG. 3 schematically depicts a blast pattern at a block 102 of a formation.
- a plurality of holes 110 are drilled in the formation.
- the holes 110 may generally be drilled in a grid pattern.
- Each hole has a diameter d and a length L.
- the hole length L may be larger or smaller than the thickness of the bench of which block 102 is a part. Any portion of a hole 110 extending beyond the bench thickness may be referred to as a sub-drill J.
- Each hole 110 may be loaded with explosives.
- the explosives may reach from the bottom of the hole to the hole opening at the surface (referred to as the collar).
- above the explosives may be one or both of a stemming region filled with material such as broken waste rock, and an open region absent of any filling material. Any stemming region has a height s and any open region has a height o.
- the volume of explosive in any hole 110 may be computed or estimated based on the diameter of the hole and the hole length L, stemming height s and open height o.
- the mass of explosive may be computed based on the volume and density of explosive.
- Holes 110 are spaced apart from one another. Horizontal separation between adjacent holes in a first direction is referred to as burden. Horizontal separation between adjacent holes in a second (e.g., perpendicular) direction is referred to as spacing.
- Drilling and explosive data may be acquired using a blast data collection device 411 (FIG. 7).
- the blast data collection device may be, for example, a laptop computer or mobile device or other suitable computing device.
- drilling and explosive data is automatically logged, e.g. by monitoring of sensors at drills or loading devices.
- the blast pattern depicted in FIG 3 is coextensive with a block 102. However, a blast pattern may cover less than an entire block, or may wholly or partially cover multiple blocks.
- FIG. 4 depicts a mining shovel 300.
- Shovel 300 includes a multi-part boom 302 with a bucket 304.
- Shovel 300 also includes a body 306 with a frame (not shown) and a cab 308, in which an operator may sit to control the shovel.
- Mining shovel 300 may be used to dig and excavate muck on a mining site, such as after the mining site has been blasted.
- the shovel 300 may be equipped with sensors for characterizing muck in a muck pile 106.
- sensors may include a machine vision system 310 including a camera capable of acquiring images of the muck pile, and an image processor for analyzing and computing sizes of fragments in the image.
- the machine vision system 310 may be configured to output one or more measurements of fragment size.
- Such a machine vision system may additionally or alternatively be installed on a haul truck (not shown) into which muck is loaded.
- the machine vision system 310 may be configured to acquire images of muck in the truck bed.
- An example of a suitable machine vision system is the FRAGTrackTM system produced by Orica Limited.
- mining shovel 300 may also include one or more positioning devices such as a GPS sensor 318.
- GPS sensor 318 may report the location of shovel 300, e.g., the shovel’s position in a mine.
- Mining shovel 300 may form part of a mine digging system 400.
- FIG. 5 depicts components of the mine digging system 400.
- Mine digging system 400 includes one or more shovel computing devices 410, each located at a respective shovel 300, one or more blast data acquisition devices 411 , data collection system 404 and operations platform 406.
- Components of digging system 400 may communicate with one another by way of one or more networks 408, which may comprise one or more local-area networks (LANs), wide-area networks (WANs), such as the Internet, or private WANs.
- networks may use any one or more of wired or wireless connection technologies, such as IEEE 802.11 (Wi-Fi), cellular, WiMax, or the like.
- mine digging system 400 may include multiple shovel computing devices 410 at multiple shovels. Additionally or alternatively, components of mine digging system 400 may be implemented fully or partially in software executed at shared hardware. For example, data collection system 404 and operations platform 406 or portions thereof may be hosted at a common server or in a common distributed (e.g., cloud) computing environment.
- cloud distributed
- FIG. 6 depicts components of a shovel computing device 410.
- Computing device 410 may, for example, be located in the cab of shovel 300.
- Computing device 410 includes memory 412, a processor 414, network interface 416, and is interconnected to sensors such as a machine vision system 310 and a GPS sensor 318.
- the processor 414 may be an intel or AMD x86-based processor, or an ARM-based processor, or any other suitable processor.
- Memory 412 (e.g., RAM) includes a computer-readable storage space accessible by processor 414 for storage of working data and code. Memory 412 also includes persistent computer-readable storage containing instructions for execution by processor 414 and for storage of data collected by machine vision system 310 and GPS sensor 318. Memory 412 may include any one or more suitable memory types, such as flash memory, hard drives or the like.
- Network interface 416 may be any suitable wired or wireless device connecting computing device 410 to a network for communication with a data host, such as data collection system 404, and operations computing device 406.
- Network interface 416 may be, for example, an Ethernet or IEEE 802.11 (Wi-Fi) network adapter.
- Network interface 416 may include an antenna for transmitting data, receiving data or both.
- the network interface may include multiple antennas, such as antennas in a phased array, and may connect to a LAN or WAN using a steered beam.
- Shovel computing device 410 is operable to receive fragmentation data from machine vision system 310 and to receive GPS coordinates from GPS 318, and to transmit fragmentation values along with associated GPS coordinates to data collection system 404, such that each set of a fragmentation value along with the associated GPS coordinates represents a measured fragmentation value acquired at that GPS position.
- fragmentation data may be used to derive a fragmentation distribution characteristic, e.g. P50.
- Data collection system 404 comprises a computing device, e.g., a server 418, and a data store 420.
- the data store may comprise any suitable type of data structure, stored on any suitable computer-readable medium.
- the data store is a database, e.g., a relational database.
- other data structures or combinations of data structures are possible.
- the data store 420 comprises fragment size data from machine vision system 310, along with location data from GPS sensor 318.
- shovel computer 302 data collection system 404 may receive sets of fragmentation values and corresponding GPS coordinates from shovel computing devices 410.
- FIG. 8A depicts an example table containing fragmentation values.
- Each record of table 101 corresponds to an image acquired by the machine vision system at a shovel or truck.
- Table 101 includes a location field 101- 1 indicating the GPS position at which the image was acquired.
- Table 101 further includes, for each of n bins into which detected fragments are sorted, a bin size value 101-2 indicating the range of fragment sizes in that bin, and a bin count value 101-3 indicating the number of detected fragments in that size range.
- Data store 420 further comprises data from the block model, such as geological and mineralogical data.
- FIG. 8B depicts an example table 103 containing block model data.
- Such block model data may include, for each block 102 in each bench 108, the unique block identifier 103-1 and a bench identifier 103- 2; and a block spatial definition 103-3
- the spatial definition may be, for example, boundaries of the block or point in the block, such as its centroid.
- the block model data may also include gross mechanical properties of the block, such as rock hardness 103-4 and blastability index 103-5, a uni-axial compressive strength 103- 6 and a bond work index (BWI) 103-7, e.g., an estimated mean BWI of the block.
- BWI bond work index
- the block model data may also include geometric properties of the block, such as a fault zone property 103-8 representing the presence (or absence) of faults in the block, and a rock quality index 103-9 representing discontinuities in the rock, measured as the percentage of chunks greater than 100 mm in length in a core sample.
- the properties may also include characterizations of the composition of the block, e.g., measurements of percentages of bornite 103-10, gypsum 103-11 , ore such as copper 103-12 and molybdenum 103-13, and an aggregate molar ratio of carbon (C), calcium (Ca), sodium (N), and potassium (K) to aluminum (Al) (“CCNK”) 103-14.
- Properties in the block model may, for example, be estimated or measured from geological core samples or geological surveys. Other geological, geotechnical and mineralogical data may also be used. For example, other measurements of rock strength, such as drop weight strength, may be used instead of or in addition to compressive strength. In some embodiments, one or more of the above-described parameters may be omitted, e.g. according to data available at a particular mine site.
- Data store 420 further comprises drilling data.
- FIG. 8C depicts an example table 105 of drilling data.
- Each record of table 105 corresponds to a drilled hole and contains properties of that hole. The properties may be measured during or after drilling of each hole and may further include one or more of estimated values and design values.
- data may be stored using one or more commercially-available products, such as CatTM MineStarTM Terrain, from Caterpillar Inc..
- table 105 includes a location value 105-1 , e.g., GPS coordinates, for each drilled hole.
- Table 105 further includes a hole depth 105-2 and a stemming length 105-3.
- a wet hole parameter 105-4 may also be recorded, indicating the presence (or absence) of water within the hole.
- the wet hole value may be obtained based on observation during drilling.
- the table may also include a drill specific energy 105-5, namely a measurement of the energy required to drill a unit volume of rock during the drilling process.
- drilling data may include additional hole dimensions, such as hole diameter.
- the hole diameter may be omitted, e.g. if holes are drilled to a standard diameter, in which case the hole diameter may be assumed as the standard value.
- hole diameter is not depicted in table 105, it may be recorded for each hole.
- Data store 420 further includes explosive data.
- the explosive data may be recorded or measured during or after loading of explosives. Additionally or alternatively, the explosive data may be based on blast design parameters. For example, the volume of explosive in a particular hole may be measured or may be assumed based on design values.
- FIG. 8D depicts an example table 107 of explosive data. Explosive data may be recorded for each hole. For example, the explosive data may characterize the amount and type of explosive loaded into each hole, along with detonation conditions for that hole.
- the explosive data may include one or more of estimated, measured, and design values, and may be obtained, for example, as part of a blast or blast design process.
- table 107 includes a location 107-1 , e.g., GPS coordinates.
- the table also includes an explosive strength value 107-2, namely, a measure of energy content per unit mass of explosive.
- a sodium nitrate explosive may have an explosive strength of approximately 650 kcal/kg and an ammonium nitrate explosive may have an explosive strength of approximately 750 kcal/kg.
- the values may further include an energy factor 107-3.
- Energy factor is a measure of how much energy is released in a blast and is the product of explosive strength and the mass of explosive.
- the values may further include a delay timing 107-4, namely the duration of any delay for detonation of explosives in that hole.
- the delay timing may be, for example: a ratio of the quantity of explosive in a hole to the time delay between detonation at that hole and at the first hole in a blast; an absolute value of delay time between detonation at a hole and at the first hole in a blast; or delay time between detonation at rows of explosives in a blast. Values may also be included for the density 107-5 of the explosive, and the detonation velocity 107-6 of the explosive.
- operations platform 406 comprises a computing device 422 including software and hardware components for computing fragmentation expected from a blast.
- operations platform 406 and data collection system 404 are shown as comprising separate computing devices, in some embodiments, operations platform 406 and data collection system 404 may share hardware.
- operations platform 406 and data collection system 404 may be implemented as applications executed on a single computing device such as a single server. Additionally or alternatively, operations platform 406 and data collection system 404 may be virtualized machines hosted at a single device.
- operations platform 406 and data collection system 404 may be executed in a cloud (e.g., distributed) computing environment, using shared hardware resources.
- cloud e.g., distributed
- computing device 422 may be a computer program or application installed on a computing device, such as a PC, phone or tablet used by a design engineer.
- operations platform 406 may be hosted on a server and accessible from another computing device, e.g., as a web application. Additionally or alternatively, functions of operations platform 406 and computing device 422 may be partially or completely automated.
- Computing device 422 comprises a plurality of application programs. As depicted, the application programs include a pre-processing module 424, blast fragmentation model 426, and a blast design module 428.
- Pre-processing module 424 is configured to combine and pre-process fragmentation data, e.g., table 101 , block model data, e.g., table 103, drilling data, e.g. table 105 and explosive data, e.g. table 107, for modelling and prediction of fragmentation and blasting optimization.
- FIG. 10 depicts an example preprocessing process 1000 carried out by pre-processing module 424.
- records in table 101 relate to individual images obtained by the machine vision system (hereinafter referred to as “image-level”). Records in table 103 relate to individual blocks 102 of the block model (hereinafter referred to as “block-level”). Records in tables 105 and 107 relate to individual holes drilled for loading with explosives (hereinafter, “hole-level”).
- Pre-processing module 424 is configured to merge such data into common reference. For example, in the depicted embodiment, tables 101, 105 and 107 are used to generate block-level fragmentation, drilling and explosive data, and to merge that data with block-level data from table 103.
- block-level data may be generated from table 101 by identifying records for images within a particular block and aggregating the distribution values from those pictures. For example, the location field 101-1 of each record of table 101 may be compared with block spatial definition 103-3. Records of table 101 may be determined to belong to a block if the location field 101-1 falls within the spatial definition field 103-3. Alternatively, if the spatial definition field 103-3 contains only a point (e.g., a centroid), records may be determined to belong to the block if the location field 101-1 is within a defined proximity to that point.
- a point e.g., a centroid
- Corresponding bin count values 101-3 from each of the records belonging to a block may be added to each other, such that the count for each bin is aggregated across the block.
- the resulting bin values may then be used to generate a cumulative distribution function, and thus to compute a fragmentation distribution characteristic (P50) value for the block.
- P50 fragmentation distribution characteristic
- Drilling and explosive data in tables 105, 107 is organized on a hole level. That is, each record corresponds to a hole of a particular pattern.
- Preprocessing module 424 is configured to collect drilling and explosive data from tables 105, 107, compute block-level values and merge the block-level values with other block-level data. Block-level drilling and explosive data may be collectively referred to as “blast features”.
- pre-processing module 424 computes block-level drilling data.
- FIG. 11A depicts an example table 105’ with block-level drilling data.
- Drilling data from table 105 may be aggregated to compute block-level values based on the location values 105-1 (FIG. 8C). For example, records may be identified as belonging to a particular block if the location values 105-1 are within the spatial definition of the block.
- the number of holes 1O5’-1 for the block may be computed as a count of the hole records that belong to that block.
- a mean hole depth 1O5’-2 and a mean stemming length 1O5’-3 may be computed from the hole records belonging to a particular block.
- a wet hole value 1O5’-4 may be computed, reflecting the presence of water in holes of that block.
- the wet hole value is the mode of wet hole values among hole records. That is, if more holes in a block are “wet” than “dry”, the block wet hole value may be “wet”.
- a wet hole value may reflect the degree to which holes are “wet”. For example, the wet hole value may be the proportion of holes in the block that are identified as wet.
- a mean drill specific energy 1O5’-5 for the block may be computed based on values from the hole records belonging to the block.
- An actual burden 1O5’-6 and a design burden 105-7’ may also be computed.
- the design burden may be defined based on recorded pattern designs.
- the actual burden may be computed based on hole locations, for example, as a mean burden between adjacent holes in the block.
- Explosive data from table 107 may also be aggregated to compute blocklevel values based on the location values 107-1 (FIG. 8D). For example, records may be identified as belonging to a particular block if the location values 107-1 are within the spatial definition of the block.
- pre-processing module 424 computes block-level explosive data.
- block-level explosive data may be stored in a table 107’.
- Block-level explosive values may be computed as the mean of values for records belonging to the block. These values may include, for example, a mean explosive strength 1O7’-1 , a mean energy factor 1O7’-2, a mean delay timing 107’- 3, a mean explosive density 1O7’-4 and a mean detonation velocity 1O7’-5.
- preprocessing module 424 merges the block-level fragmentation, drilling and explosive data with the block model data.
- FIG. 11C depicts example merged block-level data 111.
- preprocessing module 424 associates data from one or more neighbouring blocks with the merged block-level data 111.
- preprocessing module 424 computes spatial relationships between blocks based on bench number and block spatial definition. For example, the “block above” a given block may be defined as the block in the preceding bench (the bench above), with the greatest overlap with the given block. Alternatively, spatial relationships between blocks may be predetermined and recorded, e.g. in the block model.
- FIG. 11 D depicts example block above data 113 that is associated with merged block-level data 111 .
- the block above data 113 includes a rock quality index; a bond work index; a mean hole depth; a mean subdrill; a mean explosive strength; a mean stemming length and a wet hole mode.
- the blast design module 428 is operable to receive inputs defining a region to be blasted, a hole pattern, and explosives to be used. Based on those inputs, computing device 422 uses blast fragmentation model 426 to compute an expected value of a fragmentation distribution characteristic.
- the expected value may be expressed as a particle size at one or more particular points on the fragment size distribution.
- expected fragmentation distribution may be expressed in terms of a median particle size (P50), or in terms of particle sizes at first, second and third quartiles, i.e., P25, P50, P75, or the like.
- the blast fragmentation model 426 may be a model constructed using a machine learning algorithm.
- blast fragmentation model 426 may be a machine learning-based model operable to receive a set of input features related to characteristics of a blast design and of a block and to compute an estimated fragmentation distribution characteristic (e.g. P50) using those input features.
- an estimated fragmentation distribution characteristic e.g. P50
- the input features are the fields from merged block-level table 111, comprising geological and blast features and the corresponding block- above data 113 (collectively, the “block features”).
- blast fragmentation model 426 may be constructed using a machine learning algorithm.
- blast fragmentation model 426 may be a neural network, such as a multi-layer perceptron (MLP).
- MLP multi-layer perceptron
- Blast fragmentation model 426 may be trained by inputting a training data set to a MLP training algorithm.
- the training data set may include, for each block 102 at which blasting has been completed, the block features, along with the fragmentation distribution characteristic (P50) computed for that block after blasting.
- P50 fragmentation distribution characteristic
- the MLP algorithm uses the training data to create a model relating the block features to the fragmentation distribution characteristic (P50).
- the MLP may be an MLP implementation available as part of the SciKit-learn software package, accessible at https ://scikit- learn.org/stable/index. html, or an MLP implementation available as part of the Keras software package, accessible at https://keras.io/.
- the model hidden layers may be sequential dense layers with nodes [256, 256, 128, 64, 32, 8, 8, 4, 4],
- the kernel initializer may be “he_uniform” and output activation may be rectified linear unit (ReLU)
- the blast fragmentation model 426 may be used to compute a predicted fragmentation based on input block features for a block to be blasted.
- explosive features such as explosive strength, energy factor, delay timing, explosive density and detonation velocity may be plan or design values. As will be described in greater detail, these values may be defined by blast design module 428 or may be user-defined.
- Blast design module 428 is operable to produce a blast design to achieve a target fragment size distribution using blast fragmentation model 426.
- the blast design may comprise definitions of number of holes, hole geometry (depth and diameter), subdrill length, collar height, explosive type and quantity, stemming, burden and spacing of a blast pattern.
- FIG. 12 depicts a user interface 440 presented by blast design module 428.
- the user interface allows for import of a description of the geometry of a block 102 to be blasted.
- the description may be a vector file, such as a .dxf file.
- the user interface further allows for input of blast parameters including explosive type, hole diameter, subdrill length and collar height. Hole length may be determined based on block height from the imported block description.
- the blast design module defines a candidate blast design defining a set of block features based on the available merged block-level data, and block-above data for the block, and the input explosive parameters.
- blast design module may be used prior to drilling.
- the available merged block-level data and block-above data may be combined with drilling data and explosive data defined by a candidate blast design from the blast design module to define a set of block features.
- Blast design module 428 provides the set of block features to blast fragmentation model 426.
- the blast fragmentation model provides an estimated value of the fragmentation distribution characteristic P50, i.e., median fragment size, which blast design module 428 presents in a user interface.
- User interface 440 may include a dialog 442 for altering blast parameters of a candidate blast design.
- Dialog 442 may present an estimated fragmentation distribution characteristic P50 corresponding to the altered parameters.
- candidate blast designs may be iteratively tested using blast design module 428 to determine a configuration estimated to produce a desired fragmentation.
- blast design module 428 is configured to produce an optimized blast design.
- the optimized blast design may be defined as a blast design which achieves a target value for the fragmentation distribution characteristic (P50).
- the optimized design may, be optimized for low cost of blasting, where an estimated cost is based on one or both of the amount of explosive used and the amount of drilling required (e.g. number of holes). For example, total explosive quantities may be used with a per-unit weight cost for the type of explosive used, and total holes drilled may be used with a model of per-hole drilling cost to estimate cost for a particular candidate blast design.
- Blast design module 428 may use fragmentation predictions from blast fragmentation model 426 for designing blasts at blocks 102.
- FIG. 13 depicts an example process of drilling and blasting 600.
- a block in a mine is drilled and blasted.
- Parameters of the blast design are recorded and associated with corresponding blocks of a block model.
- the parameters include the number of holes per block, the burden and spacing, the diameter and length of holes, along with lengths of sub-drill and open areas, explosive type and explosive density.
- the parameters may also include measurements of explosive quantities loaded into the holes. An energy factor for the blast may be calculated based on these parameters.
- fragment sizes resulting from the blast is measured.
- a value of a fragmentation distribution characteristic is computed based on the fragment sizes. For example, a median fragment size (P50) may be computed based on machine vision measurements at a mining shovel as described above.
- Blocks 602 and 604 may be repeated one or more times to collect a data set for training a model. In an example, blocks 602 and 604 may be repeated hundreds or thousands of times. [00111] At block 606, data collected at blocks 602 and 604 are input to a machine learning algorithm as described above to produce a model for predicting fragmentation for a given block and blast design.
- a candidate blast design is defined, with parameters for a proposed blast.
- the parameters include the block geometry, type of explosive, diameter and length of holes, along with lengths of sub-drill and open areas, explosive type and explosive density.
- the parameters also include a desired value of the fragmentation distribution characteristic, e.g., a desired median fragment size (P50).
- a blast pattern is designed.
- the blast pattern may be designed by computing, using the model created at block 606, define a number of holes and a burden and spacing between the holes expected to result in the desired fragmentation.
- the process determines whether the computed blast pattern requires revision. In some examples, the determination is based on minimization of an optimization function, such minimization of a quantity of explosives required. If revision is required, the process returns to block 608 and blast parameters are modified. If revision is not required, the process advances to block 614.
- Candidate blast designs at each iteration through blocks 608-612 may be defined according to an optimization strategy.
- iteration and optimization through blocks 608-612 may be performed according to a basin hopping algorithm, for example, as described and implemented in the SciPy library, available at https://github.com/scipy/scipy.
- the designed blast pattern is implemented, and blasting is conducted. Specifically, blast holes are drilled according to the number of holes, diameter, length, burden and spacing specified by the blast pattern, and explosives are loaded into the holes as specified by the blast pattern. The explosives are detonated, and muck is collected for processing.
- the features included in the merged block level features may differ from those described above.
- the above-described features include a number of compositional features representative of mineral species present in the rock. Some or all of these features may be omitted, or additional features may be included, for compatibility of a model with a wider range of mine sites. For example, referring to FIG.
- one or more of the bornite value 103-10, gypsum value 103-11 , copper value 103-12, molybdenum value 103-13 and CCNK value 103-14 may be excluded so that a trained model could more easily be used at mine sites at which those mineral constituents are absent or are not mentioned.
- features may be commonly used across a variety of mines and mine sites.
- geotechnical features such as features relating to rock strength, geometry of discontinuities and the like may be used at many mines.
- data related to explosive quantity and type, hole geometry, quantity and spacing may be common to and available across mining sites.
- mineralogy features may vary between sites, based on prevailing rock composition at the sites.
- FIG. 15 is a representative diagram showing experimentally- determined importance of individual features to model output, i.e. relative strength of correlation to P50 values. As depicted, the most important features relate to geotechnical features such as rock strength and quality, and explosive quantity and drilling features.
- mineralogical features may therefore be omitted.
- blast fragmentation model 426 is configured to directly model the fragmentation distribution characteristic P50 from a set of input features.
- fragmentation may be modelled in two steps using the following empirical formula: (equation 2)
- x m is fragmentation, defined as median post-blast particle size (P50);
- A is the rock factor;
- K is powder factor in kg of explosive per m 3 of rock;
- Q is kg of explosives in blast holes; and
- RWS is relative weight strength of the rock.
- a rock factor A may be computed based on the P50 value obtained from the machine vision systems at shovels or trucks.
- the computed rock factor, along with block features for a particular blast, may be used as training data for the blast fragmentation model.
- a regression algorithm may be used to infer correlations between the geological features and the apparent rock factor for each historical block.
- the regression algorithm may be xgboost, described at https://qithub.com/dmlc/xqboost.
- the blast fragmentation model can receive an input set of block features and produce an estimated rock factor A.
- the estimated rock factor can then be used, along with equation 2 above to predict a fragmentation distribution characteristic value P50.
- the amount of explosive loaded into a particular hole is computed or estimated based on the hole dimensions.
- explosive loading may be directly recorded at the time of loading.
- Such measured explosive amounts may be added to the per-hole parameters described above, and per-block amounts may be computed accordingly as block features.
- hole dimension features such as hole depth, may be diminished in importance or removed entirely from the block features when directly measured explosive quantities are used.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Mining & Mineral Resources (AREA)
- Artificial Intelligence (AREA)
- General Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Computer Hardware Design (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Biophysics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Computing Systems (AREA)
- Geometry (AREA)
- Biomedical Technology (AREA)
- Geophysics (AREA)
- Remote Sensing (AREA)
- Health & Medical Sciences (AREA)
- Geochemistry & Mineralogy (AREA)
- Geology (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A mining system comprises a blast fragmentation model produced using a machine learning algorithm. The blast fragmentation model relates a fragmentation distribution characteristic to blast features based on drilling and explosive data and geological features. Values for the geological and blast features are received, and using the model a predicted value of the fragmentation distribution characteristic is generated. Blasting may be performed according to optimization by testing of multiple candidate designs.
Description
METHOD AND APPARATUS FOR DRILLING AND BLASTING WITH
FRAGMENTATION MODELING
RELATED APPLICATIONS
[0001] This claims priority from United States provisional patent application no. 63/404,414, filed September 7, 2022 and United States provisional patent application no. 63/405,284, filed September 9, 2022, the entire contents of which are incorporated herein by reference.
FIELD
[0002] This disclosure relates to mining, and more particularly to a method of drilling and blasting for excavation.
BACKGROUND
[0003] In some mining operations, rock is broken at a formation, then transported to a mill for processing. The mill may, for example, grind rock to extract mineral resources, e.g., ore such as copper ore.
[0004] Productivity of such mining operations may be significantly impacted by mill operations. For example, mill throughput may be a limiting step of the mine operation. Thus, increasing mill throughput may increase the ultimate output of a mining operation.
[0005] Mining throughput may in turn be influenced by characteristics of the input rock. For example, geological or mechanical properties of input rock, or fragment size of input rock may impact throughput. Generally, harder rock and larger fragments may be more difficult to process and may be associated with lower throughput. Softer rock and smaller fragments may be easier to process and may be associated with higher throughput.
[0006] Rock may be broken at a formation by drilling and blasting. For example, holes may be drilled in a formation, and explosives may be loaded in the holes and detonated. Generally, the quantity and strength of explosive used impacts the size of rock fragments produced from a blast. For example, the amount of explosive may depend on the number and spacing of drilled holes, and the amount of explosive in each hole. Fragment size is generally inversely related to the amount of explosive. That is, blasts involving larger amounts of explosive tend to produce smaller rock fragments. However, increasing explosive quantity also increases the cost of drilling and blasting.
SUMMARY
[0007] An example mining system comprises: a fragmentation measurement system installed at a shovel for measuring fragment size following a blast operation at a mine; a data store comprising: for each of a plurality of blocks of a mine, values for a plurality of geological features; values for a plurality of blast features based on hole drilling and explosive data for a plurality of blast events; and for individual ones of the blast events, values for a fragmentation distribution characteristic derived from measurements of fragment size resulting from the individual ones of the blast events; and a blast fragmentation model relating the geological features and blast features to the fragmentation distribution characteristic, the blast fragmentation model produced using a machine learning algorithm and operable to generate a predicted value of the fragmentation distribution characteristic from a blast defined by values for the geological features and the blast features.
[0008] An example mining method comprises: for each of a plurality of blasts at a mine: measuring hole drilling and explosive data for a blast at a mine; computing a value of a fragmentation distribution characteristic from measurements of rock fragment sizes resulting from the blast; associating the blast features with a plurality of geological features for a plurality of blocks at the mine; using a machine learning algorithm, generating a blast fragmentation model relating the fragmentation distribution characteristic to the geological features and blast features based on the hole drilling and explosive data, the blast fragmentation model operable to generate
a predicted value of the fragmentation distribution characteristic from a blast defined by values for the geological features and the blast features.
[0009] An example computing system comprises: a processor; a blast fragmentation model produced using a machine learning algorithm, relating: (i) a plurality of blast features based on hole drilling and explosive data for blast events at the mine; and (ii) a plurality of geological features for blocks at the mine; to a fragmentation distribution characteristic; computer-readable instructions which, when executed by the processor, cause the processor to receive the plurality of geological features for a block at the mine, and a corresponding set of blast features, and to generate, using the blast fragmentation model, a predicted value of the fragmentation distribution characteristic from a blast defined by the geological features for the block and the corresponding blast features.
[0010] Other aspects will be apparent from the detailed description.
BRIEF DESCRIPTION OF THE DRAWINGS
[0011] In the figures, which illustrate example embodiments:
[0012] FIG. 1 is a simplified schematic view of a region of a mine;
[0013] FIG. 2 is a schematic view of adjacent layers of a geological block model;
[0014] FIG. 3 is a schematic view of a block of the mine of FIG. 1 ;
[0015] FIG. 4 is a side view of a mining shovel ;
[0016] FIG. 5 is a schematic view of a mining data system;
[0017] FIG. 6 is a schematic view of a shovel computing device of the mining data system of FIG. 5;
[0018] FIG. 7 is a schematic view of a data collection system of the mining data system of FIG. 5;
[0019] FIG. 8A is a table showing fragmentation measurements from a machine
vision system;
[0020] FIG. 8B is a table showing properties in the block model of FIG. 2;
[0021] FIG. 8C is a table showing drilling data in the data collection system of
FIG. 7;
[0022] FIG. 8D is a table showing explosive data in the data collection system of FIG. 7;
[0023] FIG. 9 is a schematic view of an operations platform of the mining data system of FIG. 5;
[0024] FIG. 10 is a table showing a pre-processing process at the operations platform of FIG. 9;
[0025] FIG. 11 A is a table showing block-level drilling data;
[0026] FIG. 11 B is a table showing block-level explosive data;
[0027] FIG. 11C is a table showing merged block-level data defining block features;
[0028] FIG. 11 D is a table showing features of a neighbouring block associated with the merged block-level data of FIG. 11 C;
[0029] FIG. 12 is a schematic view of a user interface of a blast design module of the operations platform of FIG. 9;
[0030] FIG. 13 is a flow chart depicting a drilling and blasting method; and
[0031] FIG. 14 is a diagram depicting importance of features to output of a blast fragmentation model.
DETAILED DESCRIPTION
[0032] FIG. 1 is a schematic plan view of a region 100 of a mine. The mine may
be at a formation containing an ore, such as a copper or other metal ore. Region 100 is logically divided into a plurality of blocks 102-1 , 102-2, 102-3 (individually and collectively, blocks 102) for blasting and material removal and extraction. Mining of region 100 may be conducted stepwise. That is, segments of rock may be sequentially fragmented. Specifically, blast patterns may be defined, each pattern comprising a series of holes 104 drilled into the rock formation. Each blast pattern also defines an amount and type of explosives to be loaded into each hole and detonated (referred to as “drilling and blasting”). Each blast may produce muck 106, namely, an aggregate of fragmented rock, for removal and subsequent processing at a mill. Blast patterns may or may not be co-extensive with blocks 102. That is, a blast pattern may cover a single block, a portion of a single block, multiple blocks in their entirety, or parts of multiple blocks.
[0033] A detailed three-dimensional model, referred to as a “block model”, may be produced to represent properties of the formation to be mined. The block model defines measured and estimated geological characteristics in blocks 102 of region 100. For example, the block model may identify rock types, formation geometry, cleavage planes, mineral content, vein locations and the like.
[0034] Region 100 may be mined in horizontal slices referred to as “benches”. For example, region 100 may be horizontally divided into plurality of benches 108- 1 , 108-2, 108-3 (individually and collectively, benches 108). The vertical thickness of benches and locations and spacing of blocks 102 may be defined based on the block model, e.g., on expected locations of mineral veins, expected rock hardness and the like.
[0035] After each blast has been performed to produce muck, the muck may be removed using excavation machines referred to as shovels. Specifically, shovels dig muck and transfer the muck to vehicles for transportation. For example, rock with sufficient mineral content may be transported to a plant for processing, and waste rock may be transported for disposal. In an example, transportation is by truck. However, other modes of transport are possible.
[0036] Depending on the blast design and the geological characteristics of region 100, the muck may contain rock fragments of different sizes. The fragment sizes may be measured, and based on those measurements, rock fragmentation may be characterized, for example, in terms of fragmentation distribution, i.e. particle size distribution. For example, a muck pile may be described by particle size indices. Particle size index P50 denotes a mesh size at which 50 percent of particles pass. Likewise, particle size indices P20, P60, P80 denote mesh sizes at which 20, 60 and 80 percent of particles pass, respectively.
[0037] Fragmentation distribution may impact subsequent processing of muck. For example, muck with smaller fragment size distribution, e.g., smaller P50, may be processed more quickly, such that the throughput of a mill processing the muck may be higher.
[0038] Fragmentation of rock resulting from a blast may generally be considered as a function of rock properties, blast geometric properties, and explosive properties, i.e., fragmentation = f1 rock) 2 geometry)fi explosives) (equation 1 )
[0039] For example, rock with lower hardness and more faults or discontinuities may more easily be fragmented into small pieces, as compared to harder rock with fewer faults or discontinuities. Likewise, rock with larger proportional content of soft minerals may more easily be broken into small pieces as compared to rock with smaller proportional content of soft materials. Blasts with high explosive energy may be more likely to produce small rock fragments than blasts with lower explosive energy.
[0040] FIG. 2 schematically depicts two benches 108-1 and 108-2 in a block model. Benches 108-1 , 108-2 represent vertically adjacent layers or benches of the mine. That is, benches 108-1 , 108-2 represent rock in layers, and the rock represented by bench 108-1 is positioned immediately atop the rock represented by bench 108-2 and may be referred to as the “bench above” bench 108-2. Each bench is divided into a plurality of blocks 102. Individual blocks of bench 108-1
may be positioned directly above corresponding blocks of bench 108-2 and may be referred to as “blocks above” the corresponding blocks of bench 108-2. For example, block 102-1 of bench 108-1 is positioned directly above block 102-2 of bench 108-2.
[0041] As will be described in greater detail, the block model includes data relating to properties of each block. The properties may include, for example, location and spatial data, geometry and rock properties (e.g. mineralogical and geological properties). Such data may be relevant to fragmentation of a block resulting from a blast. For example, rock with lower hardness and more faults or discontinuities may more easily be fragmented into small pieces, as compared to harder rock with fewer faults or discontinuities. Likewise, rock with larger proportional content of soft minerals may more easily be broken into small pieces as compared to rock with smaller proportional content of soft materials.
[0042] In particular, as disclosed herein, the results of a blast given specific geometry, rock and explosive properties, may be predicted based on results of previous blasts. Such prediction may assist with efficiency of drilling, blasting and removal activities, and subsequent processing of removed material.
[0043] FIG. 3 schematically depicts a blast pattern at a block 102 of a formation. As shown, a plurality of holes 110 are drilled in the formation. The holes 110 may generally be drilled in a grid pattern. Each hole has a diameter d and a length L. The hole length L may be larger or smaller than the thickness of the bench of which block 102 is a part. Any portion of a hole 110 extending beyond the bench thickness may be referred to as a sub-drill J.
[0044] Each hole 110 may be loaded with explosives. The explosives may reach from the bottom of the hole to the hole opening at the surface (referred to as the collar). Alternatively, above the explosives may be one or both of a stemming region filled with material such as broken waste rock, and an open region absent of any filling material. Any stemming region has a height s and any open region has a height o.
[0045] The volume of explosive in any hole 110 may be computed or estimated based on the diameter of the hole and the hole length L, stemming height s and open height o. The mass of explosive may be computed based on the volume and density of explosive.
[0046] Holes 110 are spaced apart from one another. Horizontal separation between adjacent holes in a first direction is referred to as burden. Horizontal separation between adjacent holes in a second (e.g., perpendicular) direction is referred to as spacing.
[0047] Drilling and explosive data may be acquired using a blast data collection device 411 (FIG. 7). The blast data collection device may be, for example, a laptop computer or mobile device or other suitable computing device. In some embodiments, drilling and explosive data is automatically logged, e.g. by monitoring of sensors at drills or loading devices.
[0048] The blast pattern depicted in FIG 3 is coextensive with a block 102. However, a blast pattern may cover less than an entire block, or may wholly or partially cover multiple blocks.
[0049] FIG. 4 depicts a mining shovel 300. Shovel 300 includes a multi-part boom 302 with a bucket 304. Shovel 300 also includes a body 306 with a frame (not shown) and a cab 308, in which an operator may sit to control the shovel.
[0050] Mining shovel 300 may be used to dig and excavate muck on a mining site, such as after the mining site has been blasted. The shovel 300 may be equipped with sensors for characterizing muck in a muck pile 106. For example, such sensors may include a machine vision system 310 including a camera capable of acquiring images of the muck pile, and an image processor for analyzing and computing sizes of fragments in the image. The machine vision system 310 may be configured to output one or more measurements of fragment size. Such a machine vision system may additionally or alternatively be installed on a haul truck (not shown) into which muck is loaded. In such examples, the machine vision system 310 may be configured to acquire images of muck in the truck bed.
[0051] An example of a suitable machine vision system is the FRAGTrack™ system produced by Orica Limited.
[0052] In some embodiments, mining shovel 300 may also include one or more positioning devices such as a GPS sensor 318. GPS sensor 318 may report the location of shovel 300, e.g., the shovel’s position in a mine.
[0053] Mining shovel 300 may form part of a mine digging system 400. FIG. 5 depicts components of the mine digging system 400. Mine digging system 400 includes one or more shovel computing devices 410, each located at a respective shovel 300, one or more blast data acquisition devices 411 , data collection system 404 and operations platform 406. Components of digging system 400 may communicate with one another by way of one or more networks 408, which may comprise one or more local-area networks (LANs), wide-area networks (WANs), such as the Internet, or private WANs. Such networks may use any one or more of wired or wireless connection technologies, such as IEEE 802.11 (Wi-Fi), cellular, WiMax, or the like. In some embodiments, mine digging system 400 may include multiple shovel computing devices 410 at multiple shovels. Additionally or alternatively, components of mine digging system 400 may be implemented fully or partially in software executed at shared hardware. For example, data collection system 404 and operations platform 406 or portions thereof may be hosted at a common server or in a common distributed (e.g., cloud) computing environment.
[0054] FIG. 6 depicts components of a shovel computing device 410. Computing device 410 may, for example, be located in the cab of shovel 300. Computing device 410 includes memory 412, a processor 414, network interface 416, and is interconnected to sensors such as a machine vision system 310 and a GPS sensor 318.
[0055] The processor 414 may be an intel or AMD x86-based processor, or an ARM-based processor, or any other suitable processor.
[0056] Memory 412 (e.g., RAM) includes a computer-readable storage space accessible by processor 414 for storage of working data and code. Memory 412
also includes persistent computer-readable storage containing instructions for execution by processor 414 and for storage of data collected by machine vision system 310 and GPS sensor 318. Memory 412 may include any one or more suitable memory types, such as flash memory, hard drives or the like.
[0057] Network interface 416 may be any suitable wired or wireless device connecting computing device 410 to a network for communication with a data host, such as data collection system 404, and operations computing device 406.
Network interface 416 may be, for example, an Ethernet or IEEE 802.11 (Wi-Fi) network adapter. Network interface 416 may include an antenna for transmitting data, receiving data or both. In some embodiments, the network interface may include multiple antennas, such as antennas in a phased array, and may connect to a LAN or WAN using a steered beam.
[0058] Shovel computing device 410 is operable to receive fragmentation data from machine vision system 310 and to receive GPS coordinates from GPS 318, and to transmit fragmentation values along with associated GPS coordinates to data collection system 404, such that each set of a fragmentation value along with the associated GPS coordinates represents a measured fragmentation value acquired at that GPS position. As will be described in greater detail, fragmentation data may be used to derive a fragmentation distribution characteristic, e.g. P50.
[0059] Components of data collection system 404 are shown in FIG. 7. Data collection system 404 comprises a computing device, e.g., a server 418, and a data store 420. The data store may comprise any suitable type of data structure, stored on any suitable computer-readable medium. In an example, the data store is a database, e.g., a relational database. However, other data structures or combinations of data structures are possible.
[0060] The data store 420 comprises fragment size data from machine vision system 310, along with location data from GPS sensor 318. In particular, shovel computer 302 data collection system 404 may receive sets of fragmentation values and corresponding GPS coordinates from shovel computing devices 410.
[0061] FIG. 8A depicts an example table containing fragmentation values.
[0062] Each record of table 101 corresponds to an image acquired by the machine vision system at a shovel or truck. Table 101 includes a location field 101- 1 indicating the GPS position at which the image was acquired. Table 101 further includes, for each of n bins into which detected fragments are sorted, a bin size value 101-2 indicating the range of fragment sizes in that bin, and a bin count value 101-3 indicating the number of detected fragments in that size range.
[0063] Data store 420 further comprises data from the block model, such as geological and mineralogical data. FIG. 8B depicts an example table 103 containing block model data. Such block model data may include, for each block 102 in each bench 108, the unique block identifier 103-1 and a bench identifier 103- 2; and a block spatial definition 103-3 The spatial definition may be, for example, boundaries of the block or point in the block, such as its centroid. The block model data may also include gross mechanical properties of the block, such as rock hardness 103-4 and blastability index 103-5, a uni-axial compressive strength 103- 6 and a bond work index (BWI) 103-7, e.g., an estimated mean BWI of the block. The block model data may also include geometric properties of the block, such as a fault zone property 103-8 representing the presence (or absence) of faults in the block, and a rock quality index 103-9 representing discontinuities in the rock, measured as the percentage of chunks greater than 100 mm in length in a core sample. The properties may also include characterizations of the composition of the block, e.g., measurements of percentages of bornite 103-10, gypsum 103-11 , ore such as copper 103-12 and molybdenum 103-13, and an aggregate molar ratio of carbon (C), calcium (Ca), sodium (N), and potassium (K) to aluminum (Al) (“CCNK”) 103-14. Properties in the block model may, for example, be estimated or measured from geological core samples or geological surveys. Other geological, geotechnical and mineralogical data may also be used. For example, other measurements of rock strength, such as drop weight strength, may be used instead of or in addition to compressive strength. In some embodiments, one or more of the above-described parameters may be omitted, e.g. according to data available
at a particular mine site.
[0064] Data store 420 further comprises drilling data. FIG. 8C depicts an example table 105 of drilling data. Each record of table 105 corresponds to a drilled hole and contains properties of that hole. The properties may be measured during or after drilling of each hole and may further include one or more of estimated values and design values. In some embodiments, data may be stored using one or more commercially-available products, such as Cat™ MineStar™ Terrain, from Caterpillar Inc..
[0065] As depicted, table 105 includes a location value 105-1 , e.g., GPS coordinates, for each drilled hole. Table 105 further includes a hole depth 105-2 and a stemming length 105-3. A wet hole parameter 105-4 may also be recorded, indicating the presence (or absence) of water within the hole. The wet hole value may be obtained based on observation during drilling. The table may also include a drill specific energy 105-5, namely a measurement of the energy required to drill a unit volume of rock during the drilling process.
[0066] In some embodiments, drilling data may include additional hole dimensions, such as hole diameter. The hole diameter may be omitted, e.g. if holes are drilled to a standard diameter, in which case the hole diameter may be assumed as the standard value. Although hole diameter is not depicted in table 105, it may be recorded for each hole.
[0067] Data store 420 further includes explosive data. The explosive data may be recorded or measured during or after loading of explosives. Additionally or alternatively, the explosive data may be based on blast design parameters. For example, the volume of explosive in a particular hole may be measured or may be assumed based on design values.
[0068] FIG. 8D depicts an example table 107 of explosive data. Explosive data may be recorded for each hole. For example, the explosive data may characterize the amount and type of explosive loaded into each hole, along with detonation conditions for that hole. The explosive data may include one or more of estimated,
measured, and design values, and may be obtained, for example, as part of a blast or blast design process.
[0069] In the depicted example, table 107 includes a location 107-1 , e.g., GPS coordinates. The table also includes an explosive strength value 107-2, namely, a measure of energy content per unit mass of explosive. In an example, a sodium nitrate explosive may have an explosive strength of approximately 650 kcal/kg and an ammonium nitrate explosive may have an explosive strength of approximately 750 kcal/kg.
[0070] The values may further include an energy factor 107-3. Energy factor is a measure of how much energy is released in a blast and is the product of explosive strength and the mass of explosive.
[0071] The values may further include a delay timing 107-4, namely the duration of any delay for detonation of explosives in that hole. The delay timing may be, for example: a ratio of the quantity of explosive in a hole to the time delay between detonation at that hole and at the first hole in a blast; an absolute value of delay time between detonation at a hole and at the first hole in a blast; or delay time between detonation at rows of explosives in a blast. Values may also be included for the density 107-5 of the explosive, and the detonation velocity 107-6 of the explosive.
[0072] As depicted in in FIG. 9, operations platform 406 comprises a computing device 422 including software and hardware components for computing fragmentation expected from a blast. Although operations platform 406 and data collection system 404 are shown as comprising separate computing devices, in some embodiments, operations platform 406 and data collection system 404 may share hardware. For example, operations platform 406 and data collection system 404 may be implemented as applications executed on a single computing device such as a single server. Additionally or alternatively, operations platform 406 and data collection system 404 may be virtualized machines hosted at a single device. In some embodiments, operations platform 406 and data collection system 404
may be executed in a cloud (e.g., distributed) computing environment, using shared hardware resources.
[0073] For example, computing device 422 may be a computer program or application installed on a computing device, such as a PC, phone or tablet used by a design engineer. In some embodiments, operations platform 406 may be hosted on a server and accessible from another computing device, e.g., as a web application. Additionally or alternatively, functions of operations platform 406 and computing device 422 may be partially or completely automated.
[0074] Computing device 422 comprises a plurality of application programs. As depicted, the application programs include a pre-processing module 424, blast fragmentation model 426, and a blast design module 428.
[0075] Pre-processing module 424 is configured to combine and pre-process fragmentation data, e.g., table 101 , block model data, e.g., table 103, drilling data, e.g. table 105 and explosive data, e.g. table 107, for modelling and prediction of fragmentation and blasting optimization. FIG. 10 depicts an example preprocessing process 1000 carried out by pre-processing module 424.
[0076] As described above, records in table 101 relate to individual images obtained by the machine vision system (hereinafter referred to as “image-level”). Records in table 103 relate to individual blocks 102 of the block model (hereinafter referred to as “block-level”). Records in tables 105 and 107 relate to individual holes drilled for loading with explosives (hereinafter, “hole-level”). Pre-processing module 424 is configured to merge such data into common reference. For example, in the depicted embodiment, tables 101, 105 and 107 are used to generate block-level fragmentation, drilling and explosive data, and to merge that data with block-level data from table 103.
[0077] At box 1002, block-level data may be generated from table 101 by identifying records for images within a particular block and aggregating the distribution values from those pictures. For example, the location field 101-1 of each record of table 101 may be compared with block spatial definition 103-3.
Records of table 101 may be determined to belong to a block if the location field 101-1 falls within the spatial definition field 103-3. Alternatively, if the spatial definition field 103-3 contains only a point (e.g., a centroid), records may be determined to belong to the block if the location field 101-1 is within a defined proximity to that point.
[0078] Corresponding bin count values 101-3 from each of the records belonging to a block may be added to each other, such that the count for each bin is aggregated across the block. The resulting bin values may then be used to generate a cumulative distribution function, and thus to compute a fragmentation distribution characteristic (P50) value for the block.
[0079] Drilling and explosive data in tables 105, 107 is organized on a hole level. That is, each record corresponds to a hole of a particular pattern. Preprocessing module 424 is configured to collect drilling and explosive data from tables 105, 107, compute block-level values and merge the block-level values with other block-level data. Block-level drilling and explosive data may be collectively referred to as “blast features”.
[0080] At box 1004, pre-processing module 424 computes block-level drilling data. FIG. 11A depicts an example table 105’ with block-level drilling data.
[0081] Drilling data from table 105 may be aggregated to compute block-level values based on the location values 105-1 (FIG. 8C). For example, records may be identified as belonging to a particular block if the location values 105-1 are within the spatial definition of the block.
[0082] As shown in FIG. 11 A, the number of holes 1O5’-1 for the block may be computed as a count of the hole records that belong to that block. A mean hole depth 1O5’-2 and a mean stemming length 1O5’-3 may be computed from the hole records belonging to a particular block. A wet hole value 1O5’-4 may be computed, reflecting the presence of water in holes of that block. In an example, the wet hole value is the mode of wet hole values among hole records. That is, if more holes in a block are “wet” than “dry”, the block wet hole value may be “wet”. Alternatively or
additionally, a wet hole value may reflect the degree to which holes are “wet”. For example, the wet hole value may be the proportion of holes in the block that are identified as wet.
[0083] A mean drill specific energy 1O5’-5 for the block may be computed based on values from the hole records belonging to the block.
[0084] An actual burden 1O5’-6 and a design burden 105-7’ may also be computed. The design burden may be defined based on recorded pattern designs. The actual burden may be computed based on hole locations, for example, as a mean burden between adjacent holes in the block.
[0085] Explosive data from table 107 may also be aggregated to compute blocklevel values based on the location values 107-1 (FIG. 8D). For example, records may be identified as belonging to a particular block if the location values 107-1 are within the spatial definition of the block.
[0086] At box 1006, pre-processing module 424 computes block-level explosive data. As shown in FIG. 11 B, block-level explosive data may be stored in a table 107’. Block-level explosive values may be computed as the mean of values for records belonging to the block. These values may include, for example, a mean explosive strength 1O7’-1 , a mean energy factor 1O7’-2, a mean delay timing 107’- 3, a mean explosive density 1O7’-4 and a mean detonation velocity 1O7’-5.
[0087] At block 1008, preprocessing module 424 merges the block-level fragmentation, drilling and explosive data with the block model data. FIG. 11C depicts example merged block-level data 111.
[0088] At block 1010, preprocessing module 424 associates data from one or more neighbouring blocks with the merged block-level data 111. In some embodiments, preprocessing module 424 computes spatial relationships between blocks based on bench number and block spatial definition. For example, the “block above” a given block may be defined as the block in the preceding bench (the bench above), with the greatest overlap with the given block. Alternatively,
spatial relationships between blocks may be predetermined and recorded, e.g. in the block model.
[0089] In the depicted example, data related to the block above a given block is associated with its merged block-level data. FIG. 11 D depicts example block above data 113 that is associated with merged block-level data 111 . In the depicted example, the block above data 113 includes a rock quality index; a bond work index; a mean hole depth; a mean subdrill; a mean explosive strength; a mean stemming length and a wet hole mode.
[0090] As will be described in greater detail, the blast design module 428 is operable to receive inputs defining a region to be blasted, a hole pattern, and explosives to be used. Based on those inputs, computing device 422 uses blast fragmentation model 426 to compute an expected value of a fragmentation distribution characteristic. The expected value may be expressed as a particle size at one or more particular points on the fragment size distribution. For example, expected fragmentation distribution may be expressed in terms of a median particle size (P50), or in terms of particle sizes at first, second and third quartiles, i.e., P25, P50, P75, or the like.
[0091] The blast fragmentation model 426 may be a model constructed using a machine learning algorithm. In particular, blast fragmentation model 426 may be a machine learning-based model operable to receive a set of input features related to characteristics of a blast design and of a block and to compute an estimated fragmentation distribution characteristic (e.g. P50) using those input features.
[0092] In an example, the input features are the fields from merged block-level table 111, comprising geological and blast features and the corresponding block- above data 113 (collectively, the “block features”).
[0093] The block features, along with corresponding fragmentation data for completed blasts may be used as training data for blast fragmentation model 426.
[0094] Blast fragmentation model 426 may be constructed using a machine learning algorithm. In an example, blast fragmentation model 426 may be a neural network, such as a multi-layer perceptron (MLP).
[0095] Blast fragmentation model 426 may be trained by inputting a training data set to a MLP training algorithm. The training data set may include, for each block 102 at which blasting has been completed, the block features, along with the fragmentation distribution characteristic (P50) computed for that block after blasting.
[0096] Using the training data, the MLP algorithm creates a model relating the block features to the fragmentation distribution characteristic (P50).
[0097] In an example, the MLP may be an MLP implementation available as part of the SciKit-learn software package, accessible at https ://scikit- learn.org/stable/index. html, or an MLP implementation available as part of the Keras software package, accessible at https://keras.io/. The activation function used by the MLP may be an exponential linear unit (ELU), with alpha=1 .0 and dropout rate 0.05. The model hidden layers may be sequential dense layers with nodes [256, 256, 128, 64, 32, 8, 8, 4, 4], The kernel initializer may be “he_uniform” and output activation may be rectified linear unit (ReLU)
[0098] Once trained, the blast fragmentation model 426 may be used to compute a predicted fragmentation based on input block features for a block to be blasted. For such a block, at which blasting has not yet occurred, explosive features such as explosive strength, energy factor, delay timing, explosive density and detonation velocity may be plan or design values. As will be described in greater detail, these values may be defined by blast design module 428 or may be user-defined.
[0099] Blast design module 428 is operable to produce a blast design to achieve a target fragment size distribution using blast fragmentation model 426. The blast design may comprise definitions of number of holes, hole geometry (depth and diameter), subdrill length, collar height, explosive type and quantity, stemming,
burden and spacing of a blast pattern.
[00100] FIG. 12 depicts a user interface 440 presented by blast design module 428. The user interface allows for import of a description of the geometry of a block 102 to be blasted. In the depicted example, the description may be a vector file, such as a .dxf file. The user interface further allows for input of blast parameters including explosive type, hole diameter, subdrill length and collar height. Hole length may be determined based on block height from the imported block description.
[00101] As will be apparent, prior to a blast at a block, the merged block level data for that block may lack explosive data. Therefore, the blast design module defines a candidate blast design defining a set of block features based on the available merged block-level data, and block-above data for the block, and the input explosive parameters. In some embodiments, blast design module may be used prior to drilling. In such cases, the available merged block-level data and block-above data may be combined with drilling data and explosive data defined by a candidate blast design from the blast design module to define a set of block features.
[00102] Blast design module 428 provides the set of block features to blast fragmentation model 426. The blast fragmentation model provides an estimated value of the fragmentation distribution characteristic P50, i.e., median fragment size, which blast design module 428 presents in a user interface.
[00103] User interface 440 may include a dialog 442 for altering blast parameters of a candidate blast design. Dialog 442 may present an estimated fragmentation distribution characteristic P50 corresponding to the altered parameters. Thus, candidate blast designs may be iteratively tested using blast design module 428 to determine a configuration estimated to produce a desired fragmentation.
[00104] In some embodiments, blast design module 428 is configured to produce an optimized blast design. The optimized blast design may be defined as
a blast design which achieves a target value for the fragmentation distribution characteristic (P50).
[00105] Additionally or alternatively, the optimized design may, be optimized for low cost of blasting, where an estimated cost is based on one or both of the amount of explosive used and the amount of drilling required (e.g. number of holes). For example, total explosive quantities may be used with a per-unit weight cost for the type of explosive used, and total holes drilled may be used with a model of per-hole drilling cost to estimate cost for a particular candidate blast design.
[00106] Blast design module 428 may use fragmentation predictions from blast fragmentation model 426 for designing blasts at blocks 102.
[00107] FIG. 13 depicts an example process of drilling and blasting 600.
[00108] At block 602, a block in a mine is drilled and blasted. Parameters of the blast design are recorded and associated with corresponding blocks of a block model. The parameters include the number of holes per block, the burden and spacing, the diameter and length of holes, along with lengths of sub-drill and open areas, explosive type and explosive density. The parameters may also include measurements of explosive quantities loaded into the holes. An energy factor for the blast may be calculated based on these parameters.
[00109] At block 604, fragment sizes resulting from the blast is measured. At block 605, a value of a fragmentation distribution characteristic is computed based on the fragment sizes. For example, a median fragment size (P50) may be computed based on machine vision measurements at a mining shovel as described above.
[00110] Blocks 602 and 604 may be repeated one or more times to collect a data set for training a model. In an example, blocks 602 and 604 may be repeated hundreds or thousands of times.
[00111] At block 606, data collected at blocks 602 and 604 are input to a machine learning algorithm as described above to produce a model for predicting fragmentation for a given block and blast design.
[00112] At block 608, a candidate blast design is defined, with parameters for a proposed blast. The parameters include the block geometry, type of explosive, diameter and length of holes, along with lengths of sub-drill and open areas, explosive type and explosive density. The parameters also include a desired value of the fragmentation distribution characteristic, e.g., a desired median fragment size (P50).
[00113] At block 610, a blast pattern is designed. The blast pattern may be designed by computing, using the model created at block 606, define a number of holes and a burden and spacing between the holes expected to result in the desired fragmentation.
[00114] At block 612, it is determined whether the computed blast pattern requires revision. In some examples, the determination is based on minimization of an optimization function, such minimization of a quantity of explosives required. If revision is required, the process returns to block 608 and blast parameters are modified. If revision is not required, the process advances to block 614.
[00115] Candidate blast designs at each iteration through blocks 608-612 may be defined according to an optimization strategy. In some embodiments, iteration and optimization through blocks 608-612 may be performed according to a basin hopping algorithm, for example, as described and implemented in the SciPy library, available at https://github.com/scipy/scipy.
[00116] At block 614, the designed blast pattern is implemented, and blasting is conducted. Specifically, blast holes are drilled according to the number of holes, diameter, length, burden and spacing specified by the blast pattern, and explosives are loaded into the holes as specified by the blast pattern. The explosives are detonated, and muck is collected for processing.
[00117] In some embodiments, the features included in the merged block level features may differ from those described above. For example, the above-described features include a number of compositional features representative of mineral species present in the rock. Some or all of these features may be omitted, or additional features may be included, for compatibility of a model with a wider range of mine sites. For example, referring to FIG. 8B, one or more of the bornite value 103-10, gypsum value 103-11 , copper value 103-12, molybdenum value 103-13 and CCNK value 103-14 may be excluded so that a trained model could more easily be used at mine sites at which those mineral constituents are absent or are not mentioned.
[00118] As will be appreciated, some features may be commonly used across a variety of mines and mine sites. For example, geotechnical features such as features relating to rock strength, geometry of discontinuities and the like may be used at many mines. Likewise, data related to explosive quantity and type, hole geometry, quantity and spacing may be common to and available across mining sites. However, mineralogy features may vary between sites, based on prevailing rock composition at the sites.
[00119] FIG. 15 is a representative diagram showing experimentally- determined importance of individual features to model output, i.e. relative strength of correlation to P50 values. As depicted, the most important features relate to geotechnical features such as rock strength and quality, and explosive quantity and drilling features.
[00120] For improved applicability to different mining sites, mineralogical features may therefore be omitted.
[00121] As described above, blast fragmentation model 426 is configured to directly model the fragmentation distribution characteristic P50 from a set of input features. In other embodiments, fragmentation may be modelled in two steps using the following empirical formula:
(equation 2)
Where xm is fragmentation, defined as median post-blast particle size (P50); A is the rock factor; K is powder factor in kg of explosive per m3of rock; Q is kg of explosives in blast holes; and RWS is relative weight strength of the rock.
[00122] Specifically, following each completed blast, a rock factor A may be computed based on the P50 value obtained from the machine vision systems at shovels or trucks.
[00123] The computed rock factor, along with block features for a particular blast, may be used as training data for the blast fragmentation model. Specifically, a regression algorithm may be used to infer correlations between the geological features and the apparent rock factor for each historical block. In an example, the regression algorithm may be xgboost, described at https://qithub.com/dmlc/xqboost.
[00124] Once trained in this manner, the blast fragmentation model can receive an input set of block features and produce an estimated rock factor A. The estimated rock factor can then be used, along with equation 2 above to predict a fragmentation distribution characteristic value P50.
[00125] As described above, the amount of explosive loaded into a particular hole is computed or estimated based on the hole dimensions. Alternatively, explosive loading may be directly recorded at the time of loading. Such measured explosive amounts may be added to the per-hole parameters described above, and per-block amounts may be computed accordingly as block features. In some embodiments, hole dimension features, such as hole depth, may be diminished in importance or removed entirely from the block features when directly measured explosive quantities are used.
[00126] Systems and methods disclosed herein may contribute to efficient mine operations. In particular, blast patterns may be designed to produce rock fragments sized for efficient processing. In addition, such designs may limit the amount of explosive used and thus, the costs of drilling and blasting.
[00127] Of course, the above-described embodiments are intended to be illustrative only and in no way limiting. The described embodiments are susceptible to many modifications of form, arrangement of parts, details and order of operation. The invention is intended to encompass all such modification within its scope, as defined by the claims.
Claims
1 . A mining system, comprising: a fragmentation measurement system installed at a shovel for measuring fragment size following a blast operation at a mine; a data store comprising: for each of a plurality of blocks of a mine, values for a plurality of geological features; values for a plurality of blast features based on hole drilling and explosive data for a plurality of blast events; and for individual ones of said blast events, values for a fragmentation distribution characteristic derived from measurements of fragment size resulting from said individual ones of said blast events; and a blast fragmentation model relating said geological features and blast features to the fragmentation distribution characteristic, said blast fragmentation model produced using a machine learning algorithm and operable to generate a predicted value of said fragmentation distribution characteristic from a blast defined by values for said geological features and said blast features.
2. The mining system of claim 1 , comprising a blast design module for defining a blast design corresponding to a desired value of said fragmentation distribution characteristic using said blast fragmentation model.
3. The mining system of claim 2, wherein said blast design module is configured to optimize a blast by iteratively evaluating candidate blast designs using said blast fragmentation model.
The mining system of claim 3, wherein said blast design module is configured to optimize a blast for low explosive quantity at said desired fragmentation size. The mining system of claim 4, wherein said blast design module is configured to optimize a blast by iteratively evaluating candidate designs using a basin hopping algorithm. The mining system of any one of claims 1 to 5, wherein said drilling and explosive data comprises data corresponding to individual drilled holes and said blast features are block-level features corresponding to individual ones of said blocks, the system further comprising a preprocessing unit operable to compute said blast features from said drilling and explosive data. The mining system of any one of claims 1 to 6, wherein each of said blocks has an associated spatial definition and wherein at least one of said plurality of geological features for each block and said blast features for each block comprises features measured for a neighboring block positioned above that block. The mining system of claim 7, comprising a blast data acquisition system operable to record parameters of individual drilled holes and corresponding GPS locations. The mining system of any one of claims 1 to 8, wherein said blast fragmentation model is produced using a multi-layer perceptron (MLP). A mining method, comprising: for each of a plurality of blasts at a mine: measuring hole drilling and explosive data for a blast at a mine;
computing a value of a fragmentation distribution characteristic from measurements of rock fragment sizes resulting from the blast; associating said blast features with a plurality of geological features for a plurality of blocks at the mine; using a machine learning algorithm, generating a blast fragmentation model relating the fragmentation distribution characteristic to said geological features and blast features based on said hole drilling and explosive data, said blast fragmentation model operable to generate a predicted value of said fragmentation distribution characteristic from a blast defined by values for said geological features and said blast features. 1 .The mining method of claim 10, comprising defining a blast design corresponding to a desired fragmentation size using said blast fragmentation model. 2. The mining method of claim 11 , comprising optimizing a blast by iteratively evaluating candidate blast designs using said blast fragmentation model. 3. The mining method of claim 12, wherein said optimizing a blast comprises optimizing for low explosive quantity at said desired fragmentation size.
4. The mining method of claim 12 or claim 13, wherein optimizing a blast comprises iteratively evaluating candidate blast designs using a basin hopping algorithm.
5. The mining method of any one of claims 12 to 14, comprising blasting rock at the mine according to the optimized blast.
The mining method of any one of claims 11 to 15, wherein said blast features are block-level features corresponding to ones of said plurality of blocks and said hole drilling and explosive data correspond to individual drilled holes, further comprising computing said blast features from said drilling and explosive data. The mining method of any one of claims 11 to 16, wherein each one of said blocks has a spatial definition, and comprising associating individual ones of said blocks with a block positioned vertically above, wherein at least one of said geological features and said blast features for said individual ones of said blocks comprise features of the block positioned vertically above. The mining method of claim 16, comprising measuring said hole drilling and explosive data and corresponding GPS location data for individual drilled holes. A computing system, comprising: a processor; a blast fragmentation model produced using a machine learning algorithm, relating:
(i) a plurality of blast features based on hole drilling and explosive data for blast events at said mine; and
(ii) a plurality of geological features for blocks at said mine; to a fragmentation distribution characteristic; computer-readable instructions which, when executed by said processor, cause said processor to receive said plurality of geological features for a block at said mine, and a corresponding set of blast features, and to generate, using said blast fragmentation model, a predicted value of said fragmentation distribution characteristic from a blast defined by said geological features for the block and said corresponding blast features.
The computing system of claim 19, further comprising a data store comprising blast features based on hole drilling and explosive data for a plurality of blast events at a mine; and values of a fragmentation distribution characteristic from measurements of rock fragment sizes resulting from individual ones of said blast events. The computing system of claim 20, wherein said data store comprises a plurality of geological features for blocks at said mine. The computing system of any one of claims 19 to 21 , comprising computer-readable blast design instructions which, when executed by said processor, cause said processor to generate a blast design corresponding to a desired fragmentation size using said blast fragmentation model. The computing system of claim 22, wherein said blast design instructions comprise instructions which, when executed by said processor, cause said processor to optimize said blast design by iteratively evaluating candidate blast designs using said blast fragmentation model. The computing system of claim 23, wherein said blast design instructions comprise instructions which, when executed by said processor, cause said processor to optimize said blast design for low explosive quantity at said desired fragmentation size. The computing system of claim 23 or claim 24, wherein said iteratively evaluating candidate blast designs comprises using a basin hopping algorithm. The computing system of any one of claims 20 to 25, wherein each one of said blocks has a spatial definition, said blast design instructions comprising instructions, which, when executed by said processor, cause
said processor to associate individual ones of said blocks with a block positioned vertically above, wherein at least one of said geological features and said blast features for said individual ones of said blocks comprise features of the block positioned vertically above. The computing system of any one of claims 20 to 26, wherein said blast features are block-level features corresponding to ones of said plurality of blocks and said hole drilling and explosive data correspond to individual drilled holes, and said instructions comprise instructions which, when executed by said processor, cause said processor to compute said blast features from said drilling and explosive data. The mining method of claim 27, wherein said blast design instructions comprise instructions which, when executed by said processor, cause said processor to associate said drilling and explosive data with individual ones of said blocks based on measured GPS locations for said individual drilled holes.
Applications Claiming Priority (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US202263404414P | 2022-09-07 | 2022-09-07 | |
US63/404,414 | 2022-09-07 | ||
US202263405284P | 2022-09-09 | 2022-09-09 | |
US63/405,284 | 2022-09-09 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2024050637A1 true WO2024050637A1 (en) | 2024-03-14 |
Family
ID=90192628
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CA2023/051184 WO2024050637A1 (en) | 2022-09-07 | 2023-09-07 | Method and apparatus for drilling and blasting with fragmentation modeling |
Country Status (1)
Country | Link |
---|---|
WO (1) | WO2024050637A1 (en) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CA3060238A1 (en) * | 2018-10-26 | 2020-04-26 | Tata Consultancy Services Limited | Method and system for online monitoring and optimization of mining and mineral processing operations |
CA3143530A1 (en) * | 2019-08-15 | 2021-02-18 | X Development Llc | Improving blast patterns |
CN113868943A (en) * | 2021-09-15 | 2021-12-31 | 中铁十九局集团矿业投资有限公司 | Data enhancement-based blasting block degree prediction method for neural network |
-
2023
- 2023-09-07 WO PCT/CA2023/051184 patent/WO2024050637A1/en unknown
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CA3060238A1 (en) * | 2018-10-26 | 2020-04-26 | Tata Consultancy Services Limited | Method and system for online monitoring and optimization of mining and mineral processing operations |
CA3143530A1 (en) * | 2019-08-15 | 2021-02-18 | X Development Llc | Improving blast patterns |
CN113868943A (en) * | 2021-09-15 | 2021-12-31 | 中铁十九局集团矿业投资有限公司 | Data enhancement-based blasting block degree prediction method for neural network |
Non-Patent Citations (1)
Title |
---|
DUMAKOR-DUPEY NELSON K., ARYA SAMPURNA, JHA ANKIT: "Advances in Blast-Induced Impact Prediction—A Review of Machine Learning Applications", MINERALS, MOLECULAR DIVERSITY PRESERVATION INTERNATIONAL, vol. 11, no. 6, pages 601, XP093148642, ISSN: 2075-163X, DOI: 10.3390/min11060601 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Kanchibotla et al. | Modelling fines in blast fragmentation and its impact on crushing and grinding | |
Ghasemi et al. | Development of an empirical model for predicting the effects of controllable blasting parameters on flyrock distance in surface mines | |
Afradi et al. | Prediction of the penetration rate and number of consumed disc cutters of tunnel boring machines (TBMs) using artificial neural network (ANN) and support vector machine (SVM)—Case study: Beheshtabad water conveyance tunnel in iran | |
Michaud et al. | Establishing a quantitative relation between post blast fragmentation and mine productivity: a case study | |
US20200356929A1 (en) | Mining System | |
Chung et al. | Experience in fragmentation control | |
Strelec et al. | Blasting design for obtaining desired fragmentation | |
Sharma et al. | Investigation of crushed aggregate as stemming material in bench blasting: a case study | |
Bhatawdekar et al. | Building information model for drilling and blasting for tropically weathered rock. | |
WO2024050637A1 (en) | Method and apparatus for drilling and blasting with fragmentation modeling | |
Mpofu et al. | The influence of stemming practice on ground vibration and air blast | |
Ouchterlony et al. | Constructing the fragment size distribution of a bench blasting round, using the new Swebrec function | |
Nefis et al. | A model study to measure fragmentation by blasting | |
Taherkhani et al. | Transportation costs: A tool for evaluating the effect of rock mass mechanical parameters on blasting results in open pit mining | |
Jethro et al. | Effect of fragmentation on loading at Obajana cement company plc, Nigeria | |
Bye | The strategic and tactical value of a 3D geotechnical model for mining optimization, Anglo Platinum, Sandsloot open pit | |
Khorzoughi et al. | Diggability assessment in open pit mines: a review | |
Mohseni et al. | A new classification system for evaluation and prediction of unplanned dilution in cut-and-fill stoping method | |
Kansake et al. | Impact of Availability and Utilisation of Drill Rigs on Production at Kanjole Minerals Limited | |
Hoseini et al. | A quantitative model for evaluation and classification of blastings in open-pit mines | |
JP6765559B1 (en) | Rock mass size optimization system, optimization method, program, and recording medium | |
Afeni et al. | Microsoft paint imaging system–a photogrammetric approach to fragmentation measurement in rock and aggregate production | |
Singh et al. | Characterization of underground rock fragmentation | |
Yarahmadi et al. | Developing a MatLab code for determine geometry of rock mass blocks and its applications in mining and rock mechanic engineering | |
Khakulov et al. | Research and Improvement of the Current Management of the Job Rhythms of the Intellectual Complex Excavator–Dump Truck in the Adjacent Faces |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 23861757 Country of ref document: EP Kind code of ref document: A1 |